Research Topics in the Literature

down is a function of its quality which depends on the producers’ quality efforts and also on the consumers’ maintenance efforts; thus, producers should take moral hazard into account.

Chu and Chintagunta (2011) empirically examine two implications of this theory in the U.S. server and automobile market, namely that quality is negatively correlated with current warranty and quality is positively correlated with past warranties. They did not fi nd signifi cant evidence for these correlations.

Profi table Bundling

Profi table bundling, discussed in short by Choi and Ishii (2010), is the fi fth possible motive for automobile warranties; it is not considered in the rest of the literature. Firms may be able to procure cheaper repair service than consumers and practically bundle discounted pre-paid repair with the automobile. The extent of a manufacturer capacity to offer repair will vary geographically, depending mostly on its dealership network. This, combined with the fact that powertrain warranties do not vary geographically, suggests that bundling is not an important motive at least for powertrain warranties.

The structural model contains a demand model derived from a utility function and a pricing model. The demand model accounts for customer’s risk aversion to accommodate the insurance role and it incorporates customer heterogeneity to allow for the sorting motive of the warranty. The pricing model accounts for the specialty of the market: the servers are sold both directly and indirectly, and the intermediaries compete among themselves and with manufacturers.

Choi and Ishii (2010) seek empirical evidence for the role of warranties as signals of unobservable quality. They adapt the linear random utility model of consumer automobile demand to investigate the extent to which warranties affect consumer choice and the extent to which this estimated warranty effect is due to risk aversion and signalling motives. Choi and Ishii derive the marginal effects of both the binary and non-binary explanatory variables in the conditional logit and random coeffi cient logit models. Their results also demonstrate the strong role of brand loyalty, a fi nding also found by Train and Winston (2007). They also analyse the relation between warranties and brands, where the question is whether the longer warranty offsets the disadvantage in brand reputation. They compare a newer, lower reputation fi rm brand (Hyundai) to the category leader (Honda) in the small/medium car category, and the same in the luxury category (Lexus versus BMW). In both comparisons, the newer fi rm offers a two-year longer powertrain warranty. They model these relations at different levels of information asymmetry, and build up three scenarios:

– both models are unrated by Consumer Reports (CR) and there is no past purchase experience; thus, the difference in product quality is strictly the difference in brand reputation, as captured by their brand dummies;

– the category leader type is rated by CR; thus, the consumer has additional third-party information on the category leader;

– the category leader type is rated by CR and the consumer has previously purchased a car from the category leader; thus, the consumer has additional fi rst- and third-party information on the category leader.

The results of the three models contain the estimated brand dummies and the marginal effect of each additional year of powertrain warranty to the indirect utility of the consumer. Based on these estimations, Choi and Ishii calculate how many extra years of warranty should a manufacturer offer to offset the brand reputation’s disadvantage compared to the category leader.

Chu and Chintagunta (2011) examine the conditions under which each of the four competing theories on the economic roles of warranties would apply and derive testable implications from the data. Then, they assess whether these theories have empirical support in the U.S. computer server and automobile market in the context of manufacturer base warranties. The two key assumptions underlying the insurance theory is that buyers are risk-averse and the probability of product failure is non-zero. The insurance theory has the implication that the

degree of risk aversion and the warranty length should be positively correlated.

Therefore, the authors expect that the same customers will buy longer warranties when product failure increases and reliability decreases, and given a particular product failure rate, more risk-averse customers will buy longer warranties.

The key assumption of the sorting theory is the presence of consumer heterogeneity; in equilibrium, fi rms will offer a line of products distinguished by different quality, warranty and price levels. The sorting theory also implies that, in response to the menu of warranties the fi rm provides, customers with the same observable attributes (i.e. income) but different degrees of risk aversion choose different warranty contracts.

The signalling approach examines the information content of warranties under information asymmetry which is the key assumption in the sense that sellers have better knowledge about product quality than buyers. Two more trivial assumptions are that warranties are costly to the seller and the costs are systematically related to product reliability. The theory implies a positive relationship between product quality and warranty duration because only high-quality fi rms can afford long warranties because of their associated costs of fulfi lment.

The incentive role on the fi rms’ side involves two aspects: signal product quality to consumers through warranties in the presence of consumer moral hazard and motivate the fi rm to invest in product quality and supply high-quality products, at least to the extent of reducing the chances of its falling below the warranted level. Chu and Chintagunta (2011) expect a positive relationship between quality and previous warranty terms and a positive relationship between warranty terms and new product reliability.

Guajardo et al. (2012) analyse the warranty in a larger context, together with other service attributes, formulate and estimate a structural model to measure the impact of service attributes on consumer demand in the U.S. automobile industry. The authors establish three important differences compared to earlier researches. First, they focus not only on the effect of fi rms’ warranty length, but also on their service quality. Second, they study the effect on the demand of the interaction between service attributes and product quality in order to help fi rm decision-making regarding both of them. Third, they specify warranty length as endogenous in addition to price in the demand specifi cation. These authors are the fi rst to empirically analyse the value of service attributes as drivers of demand in manufacturing industries and to analyse complementarities between service attributes and product quality in the context of demand models.